Mafia 96 - Murder in Emerald City (Game Over!)

[iamausername]

To put it another way: town lynched five times, hit scum twice, and lost anyways.

To put it another way, the town's record on killing Vermillion scum was 0/8 (5 lynches, 3 vig kills). You had enough info to know that you should have been aiming for Vermillion on the last day, the fact that you couldn't manage it isn't a flaw in the setup.

And I bet I could find a whole bunch of mini games where the town lynched five times, hit scum twice, and lost anyway. Probably several where they lynched

four

times, hit scum twice, and lost anyway. (Also one where they lynched three times, hit scum three times, and lost anyway, though I won't hold that up as an example of a well-balanced setup).

[Wickedestjr]

I agree that the game was balanced, but I don't think we lynched the right people days 3 and 4.

Lol 0-for.

Blame the Vermillions for playing an awesome game.

[curiouskarmadog]

they didnt blame THAT good of a game.

really their win was a result of my claim.

[The Fonz]

To put it another way: town lynched five times, hit scum twice, and lost anyways.

Because they hit THE SAME scumgroup both times. Lynching CKD in that situation was the worst thing you could do- worse than mislynching. No one even considered the possibility of him being cerulean that last day.

All three scum killed by town were the same group. That's just unlucky. The vermillions deserved to win, because they got to the sixth night without a single one of their group lynched. Hell, if town had no lynched, there was a near-certainty of double crosskill there.

You consistently point out that town scumhunted with a decent success rate, and ignore the fact that they had no success whatsoever in hunting the group that was the greater threat. Hell, Faraday basically active lurked through, and pretty much no one called him on it at all.

[Xylthixlm]

Meh. I'm known for being power-role happy as a mod, at least compared to conventional wisdom. Then again, I recall an analysis of large themes showed that scum were winning significantly more than 50%, so "conventional wisdom" results in an underpowered town.

Part of it is lots of exposure to the IRC bot, which has an autobalancer that continually adjusts the meta to keep the town win rate at 50%. The meta it's settled on has

way

more power roles than any human would use, including me. But it gets the games to 50%.

[Xylthixlm]

To put it another way: town lynched five times, hit scum twice, and lost anyways.

Because they hit THE SAME scumgroup both times. Lynching CKD in that situation was the worst thing you could do- worse than mislynching. No one even considered the possibility of him being cerulean that last day.

All three scum killed by town were the same group. That's just unlucky. The vermillions deserved to win, because they got to the sixth night without a single one of their group lynched. Hell, if town had no lynched, there was a near-certainty of double crosskill there.

You consistently point out that town scumhunted with a decent success rate, and ignore the fact that they had no success whatsoever in hunting the group that was the greater threat. Hell, Faraday basically active lurked through, and pretty much no one called him on it at all.

We've already established that I have no ability whatsoever to selectively hunt for one scumgroup. This might explain why I consider a setup that requires selectively hunting for one scumgroup to be unfair.

[Wickedestjr]

they didnt blame THAT good of a game.

really their win was a result of my claim.

Nah, they would've won anyway.

[The Fonz]

But the games on IRC are generally smaller and significantly more night-based than the ones on here. (Besides, 50% town win across the board is overpowered. 50% is fine in single scumgroup games, but put in even an SK and it's overpowered. In two-family games, you should be aiming for something in the region of 40-30-30).

The town would be expected, given the unlynchables, to lynch at 41%. (You could argue this would be reduced by scum correctly claiming power, but then, that would just lead to town lynches and crosskills, which ultimately balances). It lynched at forty, so it's at about par. HOWEVER, the distribution of 'correct' lynches was skewed in a very unhelpful manner.

More than half the town having power roles is just shit. It isn't fun, and I couldn't give a crap if it's balanced, to be honest. So long as random lynching indicates that the town would win at least 40% and/or be the group with the single highest winning percentage, if town is continually falling short of that, then it indicates that town is playing worse than random lynching. All the invitationals followed 'conventional' wisdom, and they all ended in town wins. Which leads me to the conclusion that it's VI players, not the setups, that are dragging the win percentages down.

I often think, nowadays, that you could pretty much accurately predict the results of 80%, maybe 90 of games as an outsider looking at the setup simply by saying that the side which has the lower concentration of VIs on their side wins.

[Xylthixlm]

Meh. You're probably right.

[SerialClergyman]

Saying you lynched 2/5 mafia when the last lynch was heavily Vermillon-pushed and worse for town than noylynch/townlynch is a bit of a stretch.

Bit disappointed at the fact that Vermillon made up 4/24 town members and despite being the obvious threat, especially after the Mastin lynch D2, were killed only 1/13 times and hit all our kills - that's not regarded as playing particularly well.

We managed to win despite when it was down to 10 players all three of us were in Xyl's 'maybe not town' category yet the lynch was between a townie and our rival mafia (who only managed to save his skin via a faked bah post).

I think it's not just about the reads.. Xyl's reads were excellent, and simply stunning in hindsight, especially looking at Valhalla. But even after identifying those reads Lowell, a mason and CKD were the ones who ended up on the chopping block, and that counts for something.

[Xylthixlm]

I wonder what the bot thinks is right for a 24-player game with 2 scumteams...

Code: Select all

[mild] mafia: Godfather, Mafioso, Sibling; mafia2: Godfather, Roleblocker, Mafioso; town: Vigilante, Double Voter, Cop, Retired Wolf Hunter, Sibling, Townie (x13)
[mild] mafia: Godfather, Roleblocker, Mafioso (x2); mafia2: Roleblocker, Mafioso (x3); survivor: Survivor; town: Mason (x2), Vigilante, Redirecter, Doctor (x2), Nurse, Townie (x8)
[mild] mafia: Redirecter, Godfather, Mafioso (x2); mafia2: Godfather, Doctor, Mafioso (x2); sk: Serial Killer; town: Mason (x2), Psychiatrist, Vigilante, Cop, Doctor, Roleblocker (x5), Super-Saint, Townie (x3)

... okay, you're right.

[The Fonz]

FIVE roleblockers?

Iesus.

[Xylthixlm]

I think it's not just about the reads.. Xyl's reads were excellent, and simply stunning in hindsight, especially looking at Valhalla. But even after identifying those reads Lowell, a mason and CKD were the ones who ended up on the chopping block, and that counts for something.

The trouble is, the way I get reads, I may be right but I can't convince anyone else - I don't find the scumtells and stuff that most people expect you to use to "build a case".

The only solution is for more people to blindly follow whatever I say.

[The Fonz]

Your argument with CKD clearly begs to differ. And I thought your logical 'What SCM says and does contradict each other' case was excellent. I think it's more that the masons were hopelessly lost in this game. Oh, and Rofl was playing. And that no-one trusted you.

[SerialClergyman]

I tell you what, after this game I think I'll follow your advice.. you're probably twice as likely to be town than scum, and I'm happy with a 67% win rate

But I think it's a key issue that you raise. Look at Albert's play - the guy doesn't really post cases too often, just has gut reads that are better than chance and somehow manages to often compltely bend the town's will into lynching who he says. To be honest, I wish I had either the ability to read the game or the ability to convince, but they are definitely two seperate skills.

It does make it hard when you're a gut player to argue against your read - I pushed very hard the fact your IIoA point didn't really stand up because it was one of the few concrete things I could argue, and I felt I was on firm ground there. When you were saying that you thought I was probably scum because you had gut town reads on kmd and spyrex it was one of the most frustrating things you can imagine - how do I get out of that?

But the flip side, of course, is that if it's not easy to argue against, it's probably not too convincing to the rest of the town, who are worried and confused and frustrated and all the other things you have to deal with as a townie.

Do you think that the metagame should change to accepting other people's gut reads? Would you like to see more people wagoning/lynching because other players have a strong read on the game?

I just finished foggy londontown with iam, and in lylo iam had a sublime case, honestly the best I'd ever seen. Went through the entire game, linked every point to a post, must have taken him a week and a half. It just had one slightly wrong assumption, though, and he ended up with only half the scumteam. He was on the right track, but just couldn't quite get to the point he needed to be in time. And, because cases are much more convincing than gut reads, everyone went along with it.

[The Fonz]

SCM, you're still misrepresenting Xyl.

He didn't have a gut read on Spyre. He said that, because Spyre appeared to be having gut reads and acting on them, Spyre was likely town. I think X is doing Y, which is hard to fake as scum, is not a gut read, it's a towntell read.

[SerialClergyman]

Can't help it, just played 99 pages as scum in this game alone, going to take me a while to be reasonable again

But the point still stands - I think people are more likely to join a wagon in which you describe someone as being scum compared to one in which you say everyone else is likely town. And I also think it's easier to argue against being scum than it is to argue against everyone else being town.

It seems like a bit of a double-edged sword to me.

[The Fonz]

Yes, because to get someone else to join a wagon based on a scum read, you need to convince the other players of one read.

To get someone to join a PoE wagon, you have to convince them of three or four town reads. I really, really hate process of elimination, though, since I always seem to get fingered as scum by PoE when I'm town.

[SerialClergyman]

Do you think there's a case for using PoE to determine the scum then using a 'confirmartion bias-y' case to get them lynched?

[The Fonz]

Well, I tend to make my reads on a basis of individual things that make absolutely no sense for town, rather than 'this makes perfect sense for scum.' I tend to dislike cases based on 'well, he might have meant this by this, and done that for this motive...' what is it Xyl says? You can make ANYONE look scummy by a PbPA.

Which reminds me, I think we should add 'unreasonable burden of proof' to the logical fallacies section of the wiki. Like when Rofl was trying to get Xyl here to provide reasons why he would never possibly do the things he was doing as scum.

[The Fonz]

EBWOP: I see it's there. Though not in the complete form i would like.

[Xylthixlm]

Do you think there's a case for using PoE to determine the scum then using a 'confirmartion bias-y' case to get them lynched?

That would get

me

lynched.

[Xylthixlm]

Which reminds me, I think we should add 'unreasonable burden of proof' to the logical fallacies section of the wiki. Like when Rofl was trying to get Xyl here to provide reasons why he would never possibly do the things he was doing as scum.

That drove me nuts. I knew roflcopter was town, so I tried to actually explain my way out instead of just ignoring him like I do with most attacks. That just made me look scummier to him.

[Konowa]

Good game to read. Ticked off that I was vigged N0, but what could I do

[SpyreX]

Replacing in when I did it was killing me to watch three people I was sold on being town at each others throats.

I've learned to not second guess myself quite as much, though.